Why this poison is interesting

Insulin is the only substance in this library that appears on the other side of a poisoning as well — as an antidote. High-dose insulin euglycaemia therapy is recommended for the shocked patient poisoned by a beta-blocker or a calcium-channel blocker, where large doses of insulin support the failing myocardium. The dose used in that setting is enormous, and it reliably produces two side effects — hypoglycaemia and hypokalaemia — that clinicians spend the therapy carefully correcting.³ Insulin poisoning is those two side effects with nothing to correct them and no myocardium to rescue. It is the antidote read as a poison, which is exactly what it is.

As a poison it is unusually simple in mechanism and unusually dangerous in consequence. Insulin drives glucose into cells; an overdose drives it in faster than the liver can replace it, and the blood glucose falls until the brain — which can burn almost nothing else — begins to fail. The result is neuroglycopenia: sweating and tremor give way to confusion, focal deficits, seizures and coma, and if the hypoglycaemia is deep and prolonged, to permanent brain injury or death. There is no toxic metabolite, no organ the drug corrodes; there is only the relentless consumption of the fuel the brain cannot do without.

What makes it a modern toxicological problem rather than a simple one is the long-acting analogue. The largest review of insulin poisoning — 315 admissions from 1923 to 2022 — found that the insulin with the longest duration of action was long-acting in 83 cases and medium-acting in 116¹, and these preparations create a subcutaneous depot that releases insulin for a day or more. Glucose was infused for a median 51 hours (interquartile range 16–96 h)¹. A poisoning that a soluble insulin would settle in an evening can, from a long-acting depot, require days of glucose and vigilance — and the depot can even be surgically excised, as it was in six of those cases.¹

A poison is a drug whose kinetics have escaped its pharmacology.

The toxic principle

The toxic principle is insulin acting on its own receptor, everywhere, without regulation. Two consequences follow directly. First, glucose is driven from the blood into cells faster than hepatic glycogenolysis and gluconeogenesis can replace it, and the blood glucose falls. Second, potassium follows — insulin activates the sodium–potassium pump and drives potassium intracellularly, so the serum potassium falls alongside the glucose. Both are the ordinary actions of insulin, and both become dangerous when there is no counter-regulation and no dose ceiling.

The reason the hypoglycaemia is a catastrophe rather than an inconvenience is the brain. Neurones cannot store glucose and, over the timescale of an acute poisoning, cannot switch to another fuel; the brain is an obligate glucose consumer. As the blood glucose falls, the autonomic warning symptoms — sweating, tremor, palpitations, hunger — give way to neuroglycopenia: confusion, behavioural change, focal neurological deficits that can mimic a stroke, seizures and coma. Prolonged, deep neuroglycopenia produces permanent cortical injury. The organ that fails is the one with no metabolic alternative. Established

Toxicokinetics

The kinetics that matter are those of absorption from the injection site, because they set the duration of the poisoning — and duration is the whole clinical problem. Injected insulin is not swallowed and not metabolised to anything toxic; the question is only how long the subcutaneous depot goes on releasing hormone, and that is decided entirely by the preparation.

Insulin — the preparation, not the metabolism, sets the length of the poisoning
Parameter	Therapeutic	In overdose	Why it changes
Route and absorption	Subcutaneous injection creating a depot; absorption rate set by the preparation — rapid analogue and soluble (hours) vs long-acting analogue (a day or more)¹	The depot's release rate is the poisoning's clock	This is the only kinetic fact that matters. A soluble-insulin overdose is a problem of hours; a glargine or degludec overdose is a problem of days, because the depot keeps delivering hormone long after a normal insulin would be gone.
Distribution	Circulates and binds the insulin receptor on liver, muscle and fat; does not need to reach a specific compartment to act	Acts wherever the receptor is	There is no barrier to cross and no target compartment to reach — the hormone acts on peripheral tissues directly, which is why the effect is prompt once absorbed.
Metabolism	Degraded by insulinases in liver and kidney; no toxic metabolite	Clearance ends the action — but only of what has been absorbed	Because clearance of circulating insulin is fast, the poisoning lasts exactly as long as the depot keeps feeding it. Metabolism cannot outpace a long-acting depot, which is why the depot, not the clearance, is the problem.
Duration by preparation	Rapid analogue ~3–5 h; soluble ~6–8 h; isophane (NPH) ~12–18 h; glargine ~24 h; degludec >42 h	Ranges from an evening to several days	The single most important history in an insulin overdose is which insulin — it predicts whether the patient needs hours or days of glucose, and the review's fatalities and prolonged infusions cluster in the long-acting group.¹
Dialysability	—	No role. The injury is glucose deprivation of the brain, not a circulating small molecule to remove	There is nothing for a dialyser to achieve: the treatment is to replace what the hormone consumes — glucose — for as long as the depot lasts, not to remove the insulin.

Metabolism and the metabolites

There is no metabolite story here, and the absence is the point. Insulin is a peptide hormone degraded by insulin-degrading enzyme in the liver and kidney to inactive fragments; nothing toxic is produced, nothing is bioactivated, and the parent hormone is the entire poison. The clinically important 'metabolism' is not of the drug but of the patient: whether the liver can mobilise enough glucose to defend the blood level, which depends on glycogen stores that a starved, alcoholic or chronically ill patient may not have.

Insulin — a hormone acting without its counter-regulatory brake

Injected insulin (excess)The whole poison — no toxic metabolite is formed
Insulin receptor on liver, muscle and fat — unopposed by counter-regulationGlucose driven into cells; potassium driven into cells — the two derangements
Hypoglycaemia → neuroglycopeniaThe brain, an obligate glucose consumer, fails — seizures, coma, permanent injury
HypokalaemiaRedistributive; rebounds if over-replaced as insulin wanes
Inactive fragmentsClearance ends the action — but only of insulin already absorbed; the depot keeps feeding it¹

Elimination and accumulation

Circulating insulin is cleared quickly, so in a soluble-insulin overdose the danger ends when the injected dose is absorbed and cleared — a matter of hours. The accumulation that matters is not of the drug in the blood but of undelivered drug in the tissue depot, which continues to top up the circulation for as long as the preparation dictates. This is why the poisoning's duration is a property of the injection, not of the patient's clearance.

This is the reason surgical excision of the injection site appears in the management literature at all — in six of the 315 reviewed admissions the depot was physically removed to stop it feeding the poisoning.¹ It is a rare and drastic step, reserved for massive long-acting overdoses, and it is a striking illustration that the accumulation here is a lump of drug in the skin, not a concentration in the plasma. Inferred

Target organs — and why those

Brain

TargetNeurones deprived of glucose, their obligate fuel

Why hereThe organ that defines the poisoning and the one that suffers the permanent injury. Neurones cannot store glucose and cannot, acutely, use an alternative fuel, so a falling blood glucose starves them: autonomic warning symptoms first, then neuroglycopenia — confusion, focal deficits, seizures, coma — and, if deep and prolonged, irreversible cortical damage.¹ The brain is the target not because insulin acts on it especially, but because it is the tissue with no metabolic alternative when the glucose runs out. Established

At the bedsideSweating, tremor and hunger early; then confusion, stroke-like deficits, seizures and coma. Prolonged hypoglycaemia causes permanent neurological injury; the deaths in the review were largely neurological.¹

Whole-body potassium balance

TargetThe sodium–potassium pump, activated by insulin

Why hereIncluded because it is a diagnostic clue and a management trap. Insulin drives potassium into cells alongside glucose, producing a redistributive hypokalaemia — the same effect that makes insulin a treatment for hyperkalaemia, and the same effect seen when insulin is used as a cardiac antidote (hypokalaemia in 82% despite replacement).³ This is targeted because the pump is insulin-responsive everywhere, and the hypokalaemia is a shift, not a loss. Established

At the bedsideHypokalaemia, worst at the peak of insulin action; replace cautiously because it rebounds as the insulin wanes and potassium returns to the plasma.

The heart — indirectly

TargetNot insulin directly; the consequences of severe hypokalaemia and hypoglycaemia

Why hereIncluded because it is the mechanism by which the two derangements can kill quickly rather than slowly. Severe hypokalaemia predisposes to arrhythmia, and profound hypoglycaemia is itself arrhythmogenic and can precipitate cardiac events. The heart is a bystander organ here — injured through the electrolyte and glucose derangements rather than by any direct action of insulin on the myocardium (which, at therapeutic doses in cardiac poisoning, is actually protective). Inferred

At the bedsideArrhythmia risk from hypokalaemia and severe hypoglycaemia; corrected by treating the glucose and, cautiously, the potassium.

Timeline of effects

Insulin — a soluble overdose that settles in hours, and a long-acting one that starves for days

Time

What you seeWhat is happening

0–1 hOnset (soluble/rapid)
What you seeSweating, tremor, palpitations, hunger, anxiety — the autonomic warning of a falling glucose.
What is happeningAbsorbed insulin drives glucose into cells. With a rapid or soluble preparation the fall is quick; the warning phase may be brief or, under sedation or in hypoglycaemia unawareness, absent.
1–6 hNeuroglycopenia
What you seeConfusion, behavioural change, focal deficits, seizures, coma as the brain is starved.¹
What is happeningBlood glucose low enough that the brain cannot function. This is the dangerous phase and the one in which permanent injury is done if it is deep and unrelieved.
Hours to days (long-acting)The prolonged depot
What you seeOnset may be delayed, and hypoglycaemia recurs every time glucose is lightened — for a day or more.¹
What is happeningA glargine or degludec depot releases insulin for 24 hours or more, so the poisoning can start late and persist. The median glucose infusion in the review was 51 hours, reaching up to 96.¹ The patient is safe only when the depot is exhausted.
RecoveryResolution
What you seeRecovery is now the usual outcome — survival rose from 86% before 2000 to 96% since¹ — unless neuroglycopenia was deep and prolonged.
What is happeningOnce the depot is exhausted and euglycaemia has been maintained throughout, the poisoning is over; the residual deficit, if any, is the brain injury done at the nadir.

A gap made by a depot, not a metabolite

Insulin — the long-acting analogue depot — a subcutaneous reservoir of glargine or degludec that releases insulin for a day or more, so hypoglycaemia can be delayed in onset and persist long after a soluble insulin would have cleared

The other 39 kinds of latent phase in this library

Amphetamines and MDMA — a hormone acting normally on a kidney behaving normally, while the patient supplies the water
Anticoagulants — clotting factors that were already made, still working — the poison stops production and nothing happens until the existing stock decays
Antipsychotics — a physical object in the stomach — extended-release quetiapine forming a pharmacobezoar
Arsenic — a tissue declaring on its own timetable rather than the poison's — the arsenic is excreted within days, but the nail that was growing while it circulated does not show its white transverse line for several weeks
Arsine and stibine — a red cell mass haemolysing faster than a kidney can cope with — the exposure is over, the haemolysis is silent until the urine changes colour, and the renal failure that follows is the cause of death
Baclofen — not the poisoning but its withdrawal — an implanted pump that has silently stopped delivering, with hours to days before anything appears
Beta-blockers — a repolarisation lesion waiting for an ectopic beat to fall inside it — sotalol prolongs the QT and then, for hours, nothing happens
Calcium-channel blockers — a tablet that has not yet dissolved
Carbamazepine — reduced gut motility from its own antimuscarinic effect, forming a reservoir that delays absorption and causes relapse on the second or third day
Carbon monoxide — an inflammatory process continuing after the poison itself has gone
Chlorine and ammonia — paraquat's cause at its shortest — the time an injured alveolar epithelium takes to leak
Cocaine — a package that has not yet failed — the calcium-channel blocker's cause moved outside the tablet, where wrapping integrity cannot be measured
Colchicine — antimitotic injury to fast-dividing tissues that declares itself over 24–72 hours, after a gastrointestinal phase that can appear to settle
Death cap and the amatoxin mushrooms — the interval before amatoxin-blocked transcription starves the hepatocyte of protein
Digoxin — distribution of the drug to its target — which that page argues is not a latent phase in the paracetamol sense at all
Ethylene glycol — paracetamol's cause with a different enzyme — glycolate accumulating behind glycolate oxidase while the patient looks merely drunk
GHB and GBL — not the poisoning but its withdrawal — a half-life under an hour means admission for any reason interrupts round-the-clock redosing
Hydrofluoric acid — an ion diffusing far enough to reach a nerve ending — and the thinner the solution, the further it travels before anybody feels it
Iron — a true remission — the corrosive injury settling while absorbed iron moves into the cells it will poison, so the patient improves on their way to the dangerous phase
Ketamine — nitrous oxide's cause on a longer clock — urothelial damage accumulating over months to years of repeated exposure
Lead — bone giving lead back — a store with a half-life of 10 to 30 years releasing its contents when pregnancy, lactation, menopause or osteoporosis resorbs it, so the blood concentration rises after the exposure has ended
Lithium — transport across cell membranes
Local anaesthetic systemic toxicity (LAST) — the tissue depot — a large-volume block absorbed slowly, so systemic toxicity can appear well after the injection is finished, not only at the moment of an intravascular one
Mercury — distribution on two clocks — tissue concentrations peaking within 24 hours everywhere except the brain, which is not reached until 2 to 3 days, and which then cannot let the poison out again
Metformin — drug accumulation and a lactic acidosis that build over hours, so a well-looking early patient can deteriorate
Methaemoglobin inducers — paracetamol's cause at its shortest — one to four hours manufacturing the toxic species
Methanol — paracetamol's cause again — formate accumulating behind a folate-dependent disposal step that primates perform poorly
Methotrexate — antifolate injury to fast-dividing tissues that declares itself over days, after an early phase that can be silent
Nitrous oxide — damage accumulating to a threshold
Opioids — an antidote wearing off before the poison does — renarcotisation, the only gap in this set that treatment creates rather than reveals
Organophosphate insecticides — a fat store emptying — and, separately, a second and unexplained lesion declaring itself at a neuromuscular junction the first phase had already left
Paracetamol — time spent manufacturing a toxic metabolite
Paraquat — the body responding to an injury that is already complete
Sodium-channel blockade — a gap that cannot be shortened
Sulfonylureas — delayed and recurrent hypoglycaemia, because the drug keeps releasing insulin long after the first glucose correction
Thallium — hair on its own clock rather than the poison's — sensory symptoms come first and the alopecia that makes the diagnosis obvious follows them, well after the interval in which treatment is recommended
Theophylline — a prolonged-release formulation whose peak may not arrive for twelve hours, so the concentration rises while the patient is still being assessed
Valproate — hyperammonaemia and carnitine depletion that build after ingestion, so encephalopathy can lag a reassuring early examination
Venlafaxine and the SNRIs — the slow-release capsule — a modified-release preparation whose absorption is spread over hours, so the seizure risk of a large ingestion is still climbing long after presentation

Most latent phases in this library are metabolic — a poison must be converted or must accumulate before it acts. Insulin's gap is pharmaceutical: a long-acting analogue sits in the subcutaneous tissue releasing hormone for a day or more, so the hypoglycaemia can begin late and return again and again as treatment is lightened. It is drawn as a gap because the danger is not gone during it — it is waiting in the skin, still being absorbed.

What the mechanism predicts at the bedside

Check the glucose in every collapse, seizure or altered behaviour. Neuroglycopenia mimics stroke, psychosis, intoxication and seizure, and insulin poisoning is instantly treatable and instantly missed without a bedside glucose.²
Give glucose, and keep giving it. Intravenous dextrose — bolus then infusion targeting roughly 5.5–11 mmol/L² — is the treatment, and it may be needed for days; the review's median was 51 hours.¹ Prophylactic dextrose in an asymptomatic patient is not recommended²; treat measured hypoglycaemia.
Ask which insulin, and for how long to worry. A long-acting analogue depot lasts a day or more¹; the preparation, not the dose alone, predicts the duration of danger.
Glucagon is a first move, not the treatment. It works briefly by mobilising hepatic glycogen and fails once glycogen is depleted¹; it does not replace a glucose infusion.
Do not reach for octreotide. It suppresses endogenous insulin secretion and is the specific treatment for sulfonylurea poisoning²; it does nothing against insulin that has already been injected.
Watch and replace potassium cautiously. The hypokalaemia is redistributive³ and rebounds as the insulin wanes; over-replacement risks hyperkalaemia.
Consider excising the depot only in the extreme. Surgical removal of the injection site has been used for massive long-acting overdoses¹ — a rare, drastic measure.
Suspect a concealed source. Insulin is a classic cause of factitious hypoglycaemia; an unexplained low glucose with a suppressed C-peptide points to exogenous insulin.

The antidote, from the poison's side

There is no antidote in the sense of a drug that opposes insulin at its receptor; the treatment is replacement of what the hormone consumes. Glucose is the mainstay, given for as long as the depot lasts, and everything else is an adjunct or a caveat around it.

Intravenous glucose (dextrose): The real treatment: a bolus to correct the acute hypoglycaemia, then an infusion to maintain euglycaemia (target ~5.5–11 mmol/L) for the whole duration of insulin action — a median of 51 hours in the review, and up to several days for a long-acting depot.¹² Frequent glucose monitoring guides the rate.
Glucagon: A transient adjunct that mobilises hepatic glycogen — useful while access is obtained, useless once glycogen is depleted or in the malnourished patient.¹ Not a substitute for the glucose infusion. See glucagon.
Potassium replacement — cautious: For the redistributive hypokalaemia³, given carefully because it rebounds as the insulin wanes.
Surgical excision of the depot: Reserved for massive long-acting overdoses — physical removal of the subcutaneous reservoir to stop it feeding the poisoning, reported in six of 315 admissions.¹
Octreotide: Not indicated for insulin. It suppresses endogenous insulin release and is the treatment for sulfonylurea poisoning²; it has no effect on insulin already injected, and its appearance in a handful of reviewed cases¹ reflects its use before the distinction was clear.

Critical appraisal

There is no randomised evidence for any treatment. Moyns and Ferner identified no controlled trials and built their conclusions from a century of case reports¹; that glucose infusion, sometimes with glucagon, is almost always effective in restoring euglycaemia is a strong case-series observation, while optimum treatments to maintain euglycaemia and restore cerebral function remain uncertain.¹ The page states both the confidence and the uncertainty.
The duration figures are real but selected. The median 51-hour glucose infusion and the preparation split (83 long-acting, 116 medium)¹ come from published cases, which over-represent severe and prolonged poisonings; they describe the reported literature and the direction of the effect — longer-acting insulins, longer poisonings — rather than a population risk.
The survival improvement is genuine and cited with its statistic. Survival rose from 86% before 2000 to 96% from 2000 onward (p = 0.003)¹; it is quoted as reported, and reflects better critical care as much as any specific treatment.
The mechanistic claims are Established; the cardiac-bystander effect is Inferred. Hypoglycaemia, neuroglycopenia and redistributive hypokalaemia are standard physiology; the attribution of arrhythmia to the combined electrolyte and glucose derangements is the standard inference and is badged as such, not as a demonstrated cause of death.
Octreotide's ineffectiveness for insulin is a mechanistic argument, well grounded. It suppresses endogenous secretion², which cannot affect injected insulin; its appearance in a few reviewed cases¹ is noted as historical use rather than as evidence of benefit.
No lethal dose is given. Insulin has narrow therapeutic indices and small doses can cause hypoglycaemia², but the danger is duration and depth of neuroglycopenia rather than a threshold dose, and risk assessment belongs to TOXBASE and NPIS.

References

1
Moyns EJ, Ferner RE. Treatment of insulin poisoning: a 100-year review. Diabetic Medicine 2023 Jun;40(6):e15076. PMID 36861356. National Poisons Information Service review of 315 admissions (1923–2022): longest-acting insulin long-acting in 83 and medium-acting in 116 cases; glucose infused for a median 51 h (IQR 16–96 h); glucagon in 14, octreotide in 9; injection-site excision in 6; survival 86% pre-2000 rising to 96% from 2000 (p = 0.003); no randomised trials.
2
Klein-Schwartz W, Stassinos GL, Isbister GK. Treatment of sulfonylurea and insulin overdose. British Journal of Clinical Pharmacology 2016 Mar;81(3):496–504. PMID 26551662. Hypoglycaemia may be delayed and persistent; prophylactic dextrose not recommended; insulin-induced hypoglycaemia treated with a dextrose bolus then infusion targeting 5.5–11 mmol/L, while octreotide is the specific treatment for sulfonylurea-induced hypoglycaemia.
3
Page CB, Ryan NM, Isbister GK. The safety of high-dose insulin euglycaemia therapy in toxin-induced cardiac toxicity. Clinical Toxicology (Phila) 2018 Jun;56(6):389–396. PMID 29069937. 22 patients treated with high-dose insulin for beta-blocker and calcium-channel-blocker poisoning; 73% developed hypoglycaemia and 82% hypokalaemia despite replacement, with no apparent complications — the same two derangements that define insulin poisoning, produced deliberately as a cardiac antidote.

Insulin

At a glance